Continuous ST-segment recovery analysis and 5 static methods using ST-segment comparison between a pre- and post-treatment electrocardiogram were compared for their ability to predict infarct-related artery patency in 82 patients with acute myocardial infarction who underwent angiography a median of 124 minutes after onset of thrombolytic treatment. Accuracy at the moment of angiography was 85% (95% confidence interval [CI] 77% to 93%) for the continuous method, and 68% (CI 57% to 78%), 78% (CI 69% to 87%), 83% (CI 74% to 91%), 82% (CI 73% to 90%), and 80% (CI 71% to 89%) for the static methods. At the moment of angiography the most accurate static method and the continuous method agreed in patency assessment in 90% of the patients (CI 84% to 97%). Agreement was reduced to 83% (CI 75% to 91%) of patients when a patency assessment was performed earlier at 90 minutes after treatment onset, and was only 77% (CI 68% to 86%), at 60 minutes. Early disagreement was mainly seen when the continuous ST recording showed ST recovery from a delayed peak ST elevation after the pretreatment static electrocardiogram or when dynamic ST changes suggesting cyclic reperfusion occurred. Continuous ST-segment recovery analysis appears to be as accurate as the most accurate static methods. Continuously updated reference points appear to give important additional information when ST recovery follows a delayed peak ST elevation or when re-elevation occurs, suggesting cyclic flow changes. Such findings appear to affect about half of patients with acute myocardial infarction treated with intravenous thrombolysis, particularly early after administration of therapy.